On a typical weekday in early May, a line of roughly fifty people stretched outside the Beijing office of a leading mobile‑internet firm, each person clutching a laptop and waiting for a technician to walk them through the setup of an artificial‑intelligence assistant called OpenClaw. Similar queues appeared days later at a technology expo in Shenzhen, where engineers demonstrated how the Austrian‑originated tool could be linked to a range of Chinese services, from messaging to e‑commerce. For Sun Lei, a 41‑year‑old human‑resources manager who joined the crowd, the appeal was practical: “I’m worried about falling behind in technological developments,” she said, hoping the assistant could sift through resumes across multiple recruitment platforms with a speed no human could match.
Sun’s concern reflects a broader shift that has taken place across China’s 1.4 billion‑strong population. According to the state‑run China Internet Network Information Center, more than 600 million Chinese were using generative‑AI applications by December 2025, a jump of 142 percent over the previous year. The surge is not limited to tech‑savvy urbanites; retirees like 64‑year‑old former IT engineer Jason Tong in Shanghai have incorporated chatbots such as Doubao and Kimi into daily routines, from checking the weather to monitoring blood‑glucose levels through a Shanghai‑based health‑service that employs an AI model to generate personalized advice. “Just as carriages were eventually replaced by trains, this is bound to happen,” Tong remarked, underscoring how quickly AI has moved from novelty to necessity.
The rapid diffusion of AI tools is reshaping the competitive landscape between the United States and China. While U.S. firms still dominate raw computing power—thanks to the latest generations of GPUs from Nvidia and AMD—Chinese users are providing the data that fuels model training at an unprecedented pace. OpenRouter, a platform that monitors cross‑model traffic, reported that the weekly token consumption of Chinese AI services has recently overtaken that of their U.S. counterparts, indicating a shift from a hardware‑centric to a data‑centric advantage.
Chinese technology giants are capitalising on the momentum. Tencent has embedded OpenClaw into WeChat, the nation’s ubiquitous “super‑app,” allowing users to summon the assistant for tasks ranging from restaurant reservations to drafting legal documents. Alibaba is weaving “agentic” AI into its internal workflows, while Baidu continues to develop its own conversational models. The integration of AI into consumer‑facing products is also accelerating. Humanoid robots with sophisticated cognitive capabilities are being showcased in factories, and AI‑driven driver‑assist systems are now capable of handling complex errands such as booking a table at a restaurant.
The ecosystem approach is a point of emphasis for analysts. Lizzi Lee, a fellow at the Asia Society Policy Institute’s Center for China Analysis, noted that “the AI competition is clearly shifting from models to ecosystems,” with Chinese users acting as real‑time testers at scale. This massive, voluntary experimentation is generating feedback loops that help firms refine their products faster than in markets where regulatory or privacy constraints limit data collection.
At the grassroots level, students and entrepreneurs are already leveraging OpenClaw for commercial purposes. Zhao Yikang, a university student from Macau, used the assistant during an internship at a real‑estate agency in Zhuhai to generate promotional videos and manage social‑media accounts. When he needed a website for a photo‑service startup, the AI produced a functional site in ten minutes at a cost of less than five yuan (about 70 cents). Such low‑cost, high‑speed outcomes are prompting Chinese companies to set internal AI‑adoption targets, according to Janet Tang, a partner at AlixPartners who advises technology firms.
Government policy is reinforcing the private sector’s push. China’s five‑year plan, extended through 2030, pledges an average annual R&D spend growth of at least 7 percent, with a specific “AI plus” blueprint that calls for the integration of artificial intelligence into healthcare, education, public safety and other domains. In Shenzhen’s courts, for example, AI tools helped process 50 percent more cases last year, according to local judicial officials.
Nevertheless, the United States’ export controls on advanced semiconductors remain a structural bottleneck. Advanced AI chips from Nvidia, AMD and Intel are still subject to licensing restrictions, limiting the speed at which Chinese labs can train the largest models. Samm Sacks, a senior fellow at New America, described the restrictions as “the Achilles’ heel of many AI labs that need advanced AI chips.” Yet he also observed that the constraints have spurred greater coordination across China’s domestic design, manufacturing and adoption chain, potentially offsetting the short‑term slowdown.
Recent developments suggest that China is beginning to reduce its reliance on foreign hardware. DeepSeek, the home‑grown rival to OpenAI, unveiled a preview of its V4 model in April, noting that part of the model runs on chips produced by Huawei’s semiconductor arm. The move signals a nascent capacity to substitute U.S.‑made GPUs with domestically fabricated alternatives, a trend echoed in a Stanford Institute for Human‑Centered AI report that claimed the performance gap between the top U.S. and Chinese models has “effectively closed.”
The geopolitical stakes are evident. U.S. policymakers and leading AI firms such as Anthropic and OpenAI have accused Chinese startups of appropriating American technology, accusations that Beijing dismisses as unfounded. Lian Jye Su, chief analyst at Omdia, argues that despite the Great Firewall and export controls, the Chinese AI ecosystem is likely to narrow the performance gap further, moving from “fast follower to parallel innovator.”
For global markets and governments, China’s transformation of its massive consumer base into a live‑testing laboratory for generative AI carries profound implications. The scale of adoption provides Chinese firms with unparalleled data, while state support ensures a steady flow of talent and electricity for power‑hungry training runs. At the same time, the continued restriction on cutting‑edge chips could force Chinese developers to innovate around hardware limitations, potentially spawning new architectures that may later influence the broader AI supply chain.
In the coming years, the world will watch whether China’s ecosystem‑first strategy can sustain its rapid rollout of AI services, and how the United States will balance security concerns with the desire to stay at the forefront of a technology that is quickly becoming a cornerstone of economic productivity and geopolitical influence.
--- The article draws on reporting by AP researchers Shihuan Chen, Dake Kang in Beijing and Matt O’Brien in Providence, Rhode Island.